US7426446B2 - Calibration training for spectrometer - Google Patents

Abstract

An encoder spectrograph is used to analyze radiation from one or more samples in various configurations. The radiation is analyzed by spatially modulating the radiation after it has been dispersed by wavelength or imaged along a line. Dual encoder spectrographs may be used to encode radiation using a single modulator. A spectra sorting algorithm is used to determine calibration training spectra for a spectrometer in a spectrometer calibration training mode.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application is a division of U.S. application Ser. No. 11/169,824, filed Jun. 28, 2005, now U.S. Pat. No. 7,262,846, which claims the benefit of U.S. Provisional Application No. 60/583,834, filed Jun. 28, 2004, and U.S. Provisional Application No. 60/637,148, filed Dec. 16, 2004. Each of the foregoing is incorporated by reference in its entirety.

BACKGROUND

1. Field of the Invention

This invention relates in general to radiation spectrum analyzers and radiation image analyzers, and in particular, to radiation analyzers and encoders that use spatial modulation of radiation dispersed by wavelength.

2. Background of the Invention

Radiation spectral analysis is presently carried out in a number of ways. Dispersive and Fourier transform based analyzers are for high resolution and can be used for many different applications so that they are more versatile than existing application-specific instruments and procedures. While these analyzers offer superior spectral performance, they tend to be expensive, large, heavy and non-portable. For most applications, these instruments offer a spectral resolution that is largely unnecessary. Many analytical computations can be made using relatively few spectral measurements. The processing of the additional, unnecessary optical data reduces the speed and compromises the photometric accuracy of these instruments.

In contrast, a non-dispersive approach to spectral analysis employs a radiation source filtered by one or more bandpass to provide input to a specific analytical function. The bandpass filters are used to select one or more specific spectral components, which are characterized by a center wavelength and bandwidth. One of the principal advantages of the non-dispersive approach is the ability to individually specify the center wavelength and bandwidth of the bandpass filters to optimize the instrument for a particular application. However, if the analytical function requires a significant number of bandpass filters, the system's signal-to-noise ratio is reduced as the total energy measured in a given filter over time is inversely related to the number of filters. Furthermore, if a spectrum analyzer using this approach is configured for a first application, the filters used in the device may have to be replaced, or the number of filters changed, to adapt the analyzer to a second application. As a consequence, the non-dispersive approach has clear limitations in adaptability and in the number of spectral components that can be analyzed.

Another type of optical spectrum analyzer is the Hadamard spectrometer, which is best described as a hybrid between dispersive and non-dispersive instruments. The Hadamard spectrometer includes a spatial radiation modulator, comprising a disc made of an opaque material with slots therein that reflect or transmit radiation, where the slots have uniform transmittance or reflectance. A radiation beam is dispersed according to wavelength onto the disc and the slots are selectively spaced at different radii from the axis to form a number of different optical channels for detecting corresponding spectral components of the beam. The disc is rotated about the axis and the slots selectively encode the corresponding spectral components with a binary amplitude modulation. The encoded beam is then directed to a detector. To differentiate the intensity of the spectral component transmitted or reflected by one slot from that of another, the disc is sequentially stepped through a specific number of steps, each step comprising a binary pattern of open or closed optical channels that defines one equation in a system of simultaneous equations for the amplitudes of the spectral components. This set of simultaneous equations is then solved to yield the intensity for each channel prior to any specific analytical function.

Not only is this approach time consuming, it is prone to errors. For example, as a direct consequence of the binary encoding approach, there is no mechanism by which one can recover the actual signal levels if any one of the signal levels changes significantly over the period of rotation. It should be noted that the system of equation can be simplified if the slots are patterned such that the radiation is transmitted or blocked one spectral component at a time (e.g., a filter-wheel photometer). However, this approach changes the optical duty cycle of each of the spectral components from its optimum value of 50%, thereby degrading the signal-to-noise ratio. Finally, if a Hadamard analyzer is configured for a first application and the number of slots is changed to adapt the analyzer to a second application, the data acquisition and decoding algorithms must be changed as well. This significantly limits the instrument's adaptability.

None of the existing approaches is entirely satisfactory. Therefore, it is desirable to provide improved spectrum and image analyzers where the above-noted disadvantages are avoided or significantly diminished, and where the encoding, data acquisition and decoding are both generalized and significantly simplified.

SUMMARY OF THE INVENTION

Embodiments of the invention provide many advantages over existing radiation analyzers. In some embodiments, an encoder spectrograph employs a double-crossover optics system that allows for a highly compact encoder design. The encoder spectrograph can be made even more compact by increasing the symmetry of the optics system, for example using optics having vertices in a common plane. The encoder spectrograph may have a dual optics system that allows for encoding of two radiation beams simultaneously, thus facilitating a number of useful applications. Various additional improvements enable an encoder spectrograph having a small size and increased performance relative to previous systems.

In one embodiment, an encoder spectrograph encodes a received incident radiation beam using spatial modulation of the received beam dispersed by wavelength. A first optics disperses the received incident beam of radiation by wavelength and redirects the dispersed beam across the incident beam to form an image dispersed by wavelength along a dispersion axis. The first optics may comprise a concave diffraction grating and a crossover focusing reflector. A spatial light modulator is positioned to receive the dispersed image on a surface of the modulator. The spatial light modulator modulates the intensity of the dispersed image according to the position along the dispersion axis, thereby providing an encoded beam. A second optics collects the encoded beam from the spatial light modulator, redirects the encoded beam across the dispersed beam, and focuses the encoded beam to form a substantially overlapping image. In one embodiment, the second optics comprises a crossover fast collection optic and a detector lens. A detector element, positioned to receive the overlapping image, produces a signal representative of the encoded beam.

As described, in one embodiment, the path of the received beam crosses itself twice before reaching the detector. This double-crossover design allows the encoder spectrograph to be relatively small. In addition, the optics used to disperse, redirect, and collect or focus the beam at various stages in the optical path can be highly symmetric. The symmetry of the design further allows for a simpler and more compact encoder. Two or more of the optics components in the design may be mounted using a single mounting structure, thereby facilitating the manufacture of the encoder and the alignment of the optics in the device. In one embodiment, the diffraction grating is separately mounted, allowing for the removal of the grating and the installation of another grating, e.g., a grating designed for another application.

Other embodiments of the instrument comprise a dual encoder spectrograph that encodes separate beams using the same modulator, which may be applied in a number of applications. For example, a single instrument may employ two spectrographs operating in different spectral regions (e.g., near-infrared (NIR), mid-infrared (MIR), and/or untra-violet (UV)) but mounted onto a single encoder. In one embodiment, a sampling interface is used with two sampling sub-systems optimized for liquid (and/or solid) and gas (and/or vapor) phase samples. Each sampling sub-system can be interfaced with one of the spectrographs to enable the simultaneous analysis of samples having both liquid and solid, liquid and gas or vapor, and/or solid and gas constituents.

In another example, a single instrument employing two spectrographs is interfaced with two sampling sub-systems: an attenuated total reflectance (ATR) system for analyzing solids and/or liquids, and a gas cell for analyzing ambient air. In this manner, a first responder can analyze samples at the scene of an incident while simultaneously monitoring the ambient environment for potential hazards using a single instrument.

In another embodiment, to minimize phase error and maximize orthogonality, an in-situ phase analysis based on prime-number encoded components is used. The in-situ phases analysis may be performed using sine-only or cosine-only modulations to determine the phase of the encoded signals. In this way, the user can maximize the orthogonality of sine and cosine components having the same modulation frequency.

In another embodiment, the encoder spectrograph analyzes a continuous stream of spectra acquired during a calibration training mode. The spectrograph-encoder analyzer automatically logs those unique spectra required to build a calibration model, also called calibration training spectra. This enables a user to identify and quantify analytes by analyzing unknown spectra by acquiring the proper reference spectra required for the underlying calibration. In another embodiment, a calibration training spectra can be determined using a spectra sorting algorithm. By calculating a residual score for a new spectrum for each of the existing calibration training spectra, the closest match to the new spectrum can be identified by identifying the minimum residual score.

In another embodiment, a user can “tag” spectra with analyte species and concentration information, possibly during the collection of the calibration training spectra. One or more selected unique calibration training spectra can be compared with spectra from a spectral library or other database for further verification of assigned tags. In this way, users can share information and compare results between two or more instruments.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic side view of an optical system for an encoder spectrograph, in accordance with an embodiment of the invention.

FIG. 2 is a schematic top view of the optical system for the encoder spectrograph shown inFIG. 1, in accordance with an embodiment of the invention.

FIG. 3 is a schematic perspective view of the optical system for the encoder spectrograph shown inFIG. 1, in accordance with an embodiment of the invention.

FIG. 4 is an assembly view of a double-crossover encoder spectrograph, in accordance with an embodiment of the invention.

FIG. 5 is a perspective view of the double-crossover encoder spectrograph shown inFIG. 4, in accordance with an embodiment of the invention.

FIG. 6 is a plan view of the double-crossover encoder spectrograph shown inFIG. 4, in accordance with an embodiment of the invention.

FIG. 7 illustrates a method of determining calibration training spectra for a spectrometer, in accordance with an embodiment of the invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

Disclosed herein are embodiments of an encoder spectrograph and related systems, subsystems, and applications for analyzing radiation from a sample. In various embodiments, an encoder spectrograph uses a spatial light modulator to encode dispersed radiation. The spatial light modulator for the encoder spectrograph may be implemented by a reflective modulator disc having a pattern of radiation intensity filters. Generally, embodiments of radiation analyzers and encoders that use spatial modulation of radiation dispersed by wavelength are described in U.S. Pat. Nos. 6,271,917, 6,999,165, and 6,995,840, each of which is incorporated by reference in its entirety. In one embodiment in accordance with the present invention, an encoder spectrograph comprises a Littrow-mount double-crossover optical system, which provides a number of advantages over previous designs.

Optical System for Encoder Spectrograph

FIG. 1 is a side view (i.e., in the global XZ-plane) of an optical system for anencoder spectrograph100, in accordance with one embodiment of the invention. The optical system comprises anentrance aperture101, aconcave diffraction grating102, acrossover focusing reflector103, a spatiallight modulator104, a crossoverfast collection optic105, adetector lens106, adetector window107, and adetector element108. Additional elements can be included to improve performance or to accommodate different radiation sources, sample accessories, spatial light modulators, and/or radiation detectors. As shown inFIG. 1, theentrance aperture101 is located in a plane substantially parallel to the global XY-plane, and the normal to theentrance aperture101 and the normal to thedetector element108 are parallel.

As depicted inFIG. 1, an incident radiation beam B₀is launched throughentrance aperture101 along the Z-axis and illuminatesconcave diffraction grating102. The diffracted radiation beam B_difis directed towardcrossover focusing reflector103, which redirects and focuses B_difto form dispersed image I_dison the surface of a spatiallight modulator104. The dispersed image I_disis encoded by theencoder104 to provide an encoded radiation beam B_enc. The encoded radiation beam B_encis collected, focused and redirected by crossoverfast collection optic105 toward adetector lens106. Thedetector lens106 provides additional focusing to project the encoded radiation beam B_encthrough adetector lens107 and form a substantially overlapping image I_deton the surface ofdetector element108, as described in U.S. Pat. No. 6,995,840.

In one embodiment, theconcave diffraction grating102 has an illuminated diameter of 34 mm and a radius of curvature R₀of approximately 100 mm, resulting in an F/# of approximately 1.5. Other parameters can be chosen for theconcave diffraction grating102 for various applications, within the scope of the invention. Theconcave diffraction grating102 may have variable line (groove) spacing or holographic etching to minimize the number of optical elements. In the embodiment shown inFIG. 1, theconcave diffraction grating102 has a biconic substrate and a variable line spacing function that is modeled as a polynomial in both the local X and Y-axes. Other embodiments incorporating adiffraction grating102 comprising torroidal, aspherical, spherical or plane mirror surfaces with ruled or holographic etched patterns are within the scope of the invention.

As shown, theconcave diffraction grating102 is rotated relative to the plane ofentrance aperture101. The rotation is about an axis through its center parallel to the Y-axis (i.e., inclination), and also about an axis through its center parallel to the X-axis (i.e., auto collimation, shown inFIG. 2).

The radiation diffracted by the diffraction grating102 (diffracted beam B_dif) is directed toward acrossover focusing reflector103, preferably at an inclination angle θ_inc. From a sequential ray tracing perspective, the value for inclination angle θ_incthat minimizes aberration in the dispersed image is zero, but that configuration would result in an unphysical design where thecrossover focusing reflector103 obstructs incident radiation beam B₀. In one embodiment, therefore, the value for inclination angle θ_incis optimized to minimize aberrations and maximize throughput. The inclination angle θ_incmay be optimized in one way by selecting the inclination angle θ_incthat just clears thecrossover focusing reflector103 out of the incident beam path. In one example, this inclination angle θ_incis approximately 7 to 8 degrees.

Thecrossover focusing reflector103 directs the diffracted radiation beam B_difacross incident radiation beam B₀to form a dispersed image I_dison the surface of a reflective spatiallight modulator104. Thereflector103 also provides additional focusing power to enableconcave grating102 to have a substrate shape that minimizes aberration in the dispersed image and maximized system throughput. In one embodiment, thereflector103 has a cylindrical surface with an axis in the XZ-plane and a radius of curvature of roughly 4R₀to provide additional focusing power along the dispersion axis. Other parameters may be selected in view of the trade off between the radius of curvature of the focusingreflector103 and the corresponding radius of curvature of thediffraction grating102, due to a reduction in the magnitude of the diffraction angles at the extremes of the dispersed image. For example, thecrossover focusing reflector103 may be modeled as a cylindrical surface with axis in the YZ-plane and as a biconic or torroid. Other embodiments may incorporate acrossover focusing reflector103 that comprises aspherical, spherical, or planar mirror surfaces.

The spatiallight modulator104 is located in a plane that is substantially parallel to the global YZ-plane, as defined inFIGS.1 and 2. In one embodiment, the spatiallight modulator104 for theencoder spectrograph100 may comprise a rotating encoder disc patterned, as described in U.S. Pat. No. 6,995,840. In another embodiment, the spatiallight modulator104 may comprise other types of spatial light modulators, including addressable LCDs and MEMS devices. In addition,optical elements101 through103 may also be used with a conventional detector array.

In one embodiment, the encoder pattern is part of the optimized optical system, and the radial position and radial width of the annular regions encompassing the radiation filters on the spatiallight modulator104 can be optimized along with the other optical elements to maximize performance. Because of the flexibility in specifying the size, shape, and location of the radiation filters, the spatiallight modulator104 can be optimized to compensate for artifacts introduced by holographic grating manufacturing methods, e.g., non-parallel grooves. For a given holographic grating, an encoder pattern can be designed to improve system performance significantly (e.g., the details of the radiation filters or the configuration of LCDs or MEMS elements).

Preferably, the holographic grating and the encoder pattern are designed in consideration of the other to exploit annular shaped spectral sub-images in the dispersed image, which in turn minimizes other aberrations or throughput limitations in the system.

In the embodiment shown, the diffracted radiation beam B_difhas an angle of incidence in the XZ-plane of roughly 15 degrees with respect to the surface of themodulator104. This angle of incidence enables the positioning of acollection optic105. Themodulator104 encodes the dispersed image I_disby modulating reflectance to provide an encoded beam B_enc. The encoded beam B_encis collected by thecollection optic105 and directed across the diffracted radiation beam B_difonto the surface of adetector lens106. In one embodiment, thecollection optic105 is modeled as a biconic, an asphere, or as a torroid surface (e.g., using ZEMAX® optical design software); however, other types of optics can be used to collect the encoded beam B_enc. Other embodiments may incorporate acollection optic105 that comprises aspherical, spherical, or planar mirror surfaces.

Encoded radiation beam B_encis focused by thecollection optic105 and thedetector lens106 through adetector window107 and onto the surface of adetector element108 to form detector image I_det. Although the surface of thedetector lens106 may be strictly spherical to minimize cost, the aperture of thedetector lens106 can be cropped to prevent a non-illuminated portion of the lens from obstructing incident radiation beam B₀. To crop thedetector lens106, it may be notched, cut, or ground flat along an edge. The cropping also serves to prevent mechanical interference with spatiallight modulator104. In the embodiment shown inFIG. 1, thedetector lens106 is a plano-convex design with a spherical surface. Other embodiments may incorporate adetector lens106 that has one or more biconic, torroidal, or aspherical surfaces (convex or concave).

FIG. 2 shows a top view of theencoder spectrograph100 ofFIG. 1, illustrating the relationship between the auto collimation angle θ_incand the dispersed image I_dison the surface of themodulator104.FIG. 2 shows a high degree of beam symmetry—approximate mirror symmetry—through the XZ-plane, which simplifies the placement and fabrication of the optical components (e.g., thelens106 is spherically symmetric). As shown inFIG. 2, the centers of the optical elements are each substantially in the global XY-plane.

FIG. 3 is a perspective view of the prototype system described inFIG. 1, illustrating the three-dimensional relationship between the optical components and the crossing beam paths. Preferably, the auto collimation angle θ_incand the grating pattern (e.g., holographic etching or ruled lines/grooves) are optimized (1) to project the correct spectral range over the surface of themodulator104, (2) to maximize spectral purity in the encoded components, and (3) to maximize the collection efficiency ofpost-encoder optics105 through108. In one embodiment, the dispersed image spans about the 3 to 5-micron spectral range, the nominal line frequency of grating102 is about 110 g/mm, and the auto collimation angle is about 13-14 degrees.

In one embodiment, the optical surfaces of the optical elements are optimized to maximize the spectral resolution of the encoded dispersed image I_disand to maximize the overlap of the encoded spectral components ondetector element108 to produce detector image I_det. In various embodiments, the optics system for theencoder spectrograph100 is optimized using the merit function design strategies, as detailed in U.S. Pat. No. 6,995,840.

In an example embodiment, the dispersed image I_disis roughly 2 mm in width and 25.4 mm in length, thedetector element108 is 3 mm by 3 mm, and the optical path between the modulator104 and thedetector element108 is roughly 40 mm. The optical system of theencoder spectrograph100 occupies a space of roughly 110 mm by 50 mm by 60 mm, along the Z, Y, and X-axes, respectively. The radius of theencoder disc104 is slightly greater than the focal length of the grating102, thereby enabling a very efficient use of space, e.g., in and around the spindle motor and encoder disc. The minimum spectral efficiency is greater than 80% at the encoder and greater than 70% for 128 spectral components encoded in the 3 to 5-micron spectral range.

As described in U.S. Pat. No. 6,995,840, the distance of the optical path between the reflective spatiallight modulator104 and thedetector element108 is preferably designed to be as short as possible for a given detector element size (e.g., preferably less than 45 mm for a 3 mm by 3 mm detector element). A post-optic, defined byoptical components105 through108, is a variation of and can be replaced by a short-path post encoder optics described in U.S. Pat. No. 6,995,840. One advantage of the post-collection optic shown inFIGS. 1-3 is the elimination of one of the reflective surfaces described in U.S. Pat. No. 6,995,840. Thefast collection optic105 can be modeled as a biconic, a torroid, and/or as an apertured asphere. In one embodiment, a biconic with additional aspherical terms added to the local y-radius of curvature is used.

In one embodiment, theencoder spectrograph100 is designed by approximating the shape of the intercept of the annular radiation filters and the dispersed image as rectangular bins to allow the axis of the dispersed image to be aligned parallel or anti-parallel to the radius of an encoder disc. This feature enables the user to reverse the order of wavelength versus encoder radius by moving the spectrograph to the opposite side of the encoder disc. Since the modulation frequencies generally increase with increasing encoder disc radius, this flexibility allows a designer to select whether the highest modulation frequencies correspond to the longest or the shortest wavelengths. This in turn allows the designer to minimize the effects from residual interferences from imperfect orthogonal encoding functions (e.g., integer harmonics) by encoding critical spectral regions at lower frequencies. In the 3 to 5-micron spectral range, for example, one could encode the CO₂spectral region at higher frequencies to minimize the residual interference effects of dynamically varying CO₂concentrations on the encoded hydrocarbon spectral components.

The use of rectangular-shaped bins also enables an instrument incorporating a single encoder and twin spectrographs with entrance apertures on the same side of the housing. For example, the each spectrograph can be coupled to its own sample cell located on the same side of the instrument. Different length sample cells can be used to maximize sensitivity and minimize absorption saturation for applications involving chemical mixtures that have a wide range of constituent concentrations, e.g., natural gas. This enables the real-time spectral analysis of a single in-line sample path with two different length sample cells and/or two distinct spectral ranges.

Theencoder spectrograph100 can be designed such that the centers of the individual optical components are located in a substantially common plane parallel to a perpendicular bisector of the midpoint of the dispersed image I_dis. Theencoder spectrograph100 may also be designed to minimize its footprint, particularly along the dispersion axis. The use of a double-crossover beam path and component symmetries help to reduce the encoder's footprint. The position of theconcave diffraction grating102 may located behind the position of encoder spindle motor to enable the dispersed image I_disto be located as close to the rotational axis as possible. In this manner, instrument designs employing smaller encoder discs are possible.

In one embodiment, thegrating102 is mounted in an inclined Littrow auto-collimation configuration so that the diffraction angles are symmetrical about the midpoint of the dispersed image I_dis. This helps to minimize aberrations or distortions to the dispersed image I_disand to maximize the collection efficiency of the post-encoder optics (components105 through108). The Littrow auto-collimation mounting of the grating102 makes theencoder spectrograph100 highly symmetrical about the midpoint of the dispersed image I_disand thereby enables the grating102 to be located on either side of a spindle motor used to rotate theencoder104.

The inclination angle θ_incof thediffraction grating102 may be defined as a rotation about an axis that passes through the origin of thediffraction grating102 that is parallel to the nominal dispersion axis (e.g., the radial axis of the encoder disc104). The optical throughput of theencoder spectrograph100 is normally constrained by the inclination angle θ_incof theconcave diffraction grating102 and the corresponding placements of thecrossover focusing reflector103 and thedetector lens106. These components are preferably arranged to optimize the tradeoff between throughput and aberrations in the dispersed image. The inclination angle θ_incof thediffraction grating102 can be optimized to maximize throughput and minimize aberrations. Preferably, thecrossover focusing reflector103 anddetector lens106 are notched or otherwise shaped to maximize optical throughput.

In one embodiment, theencoder spectrograph100 includes a single bracket to hold and position thecrossover focusing reflector103, thefast collection optic105, thedetector lens106, thedetector window107, and thedetector element108. This simplifies assembly of the system and facilitates alignment of the optical components. In addition, the single bracket construction provides a more stable and thermally predictable alignment. Preferably, thecrossover focusing reflector103, thefast collection optic105, and the mounting provisions for thelens106 are combined into a single injection-molded optic. To create the reflective metallic surfaces for thereflectors103 and105, a metal may be evaporated on the injection-molded optic by placing it on a rotating chuck in an evaporation chamber.

In one embodiment, the entrance axis of the encoder spectrograph100 (i.e., the normal to the entrance aperture101) and the detector axis (i.e., the normal to the surface of the detector element108) are parallel to each other and to the plane of the spatiallight modulator104. In addition, theentrance aperture101 is perpendicular to the side of the enclosure. This configuration enables the use of an enclosure with “straight-in” optical coupling to simplify the integration and alignment of sampling accessories. Alternately, folding mirrors can be introduced to enable other configurations for the entrance aperture and the detector plane.

Non-lineal and nonlinear models for the dispersed image may be introduced to minimize optical aberrations and maximize throughput. Because of the flexibility in defining the annular regions encompassing the encoding tracks, the radial position and the radial width of the encoding tracks can be varied to compensate for nonlinearities (e.g., compressions and expansions along the radial axis). The phases of the encoding patterns or the decoding algorithm can be varied to compensate for variations in the position of the spectral centroids along the azimuthal axis within an annular region.

Holographic Grating

Because of the effects of the non-zero inclination angle θ_inc, illustrated inFIG. 1, improved efficiency and spectral purity may be obtained when the holographic recording coordinates are located out of a plane containing the surface normal at the center ofgrating substrate102. In one embodiment, this is facilitated by creating a ZEMAX® compatible User Defined Surface that constrained ZEMAX® Hologram Type 1 coordinates to lie in a plane that is parallel to the Y-axis ofFIG. 1. The highest efficiency and spectral purity may be obtained when the holographic recording coordinates are located in the plane shown inFIG. 2 that makes an angle θ_increlative to the surface normal at the center ofgrating substrate102.

In a standard detector array instrument, the pitch between detector elements is fixed and is typically uniform. Spectrograph designs employing detector arrays are therefore optimized to provide a substantially linear relationship between wavelength and the distance between detector elements along the detector array axis (i.e., a linear dispersion function). In an embodiment of the invention, improved efficiency and spectral purity are obtained by incorporating radiation filters that have non-equal widths and are centered at non-equal intervals along the encoding axis of themodulator104, the Y-axis inFIG. 2. This corresponds to a non-linear relationship between wavelength and the distance between encoding tracks along the encoding axis (i.e., a non-linear dispersion function). In addition, the efficiency and spectral purity of the system can be optimized for constant energy separation and resolution, as well as arbitrary configurations of spectral components.

In one embodiment, the optical system is optimized using a multi-order polynomial in the position along the Y-axis ofFIG. 2 to define the position and widths of the annular regions bounding the radiation filters. The highest efficiency and spectral purity may be obtained when the holographic recording coordinates are optimized along (simultaneously) with the multi-order polynomial coefficients defining the position and widths of the encoding tracks (encoding bins) along the Y-axis ofFIG. 2.

In another embodiment, thediffraction grating102 is optimized to exploit the true annular-shaped intercept between the dispersed image I_disand the concentric radiation filters on theencoder disc104. The overlap of an annular region of radial with w₀located at radius r₀with a rectangle of width w₀and length l₀is only 100% in the limit of infinite radius r₀. Therefore, creating dispersed images where spectral components occupy annular bins enables the use of an encoder disc with a smaller radius without a significant loss in spectral resolution.

Practical Embodiments for Optical Elements

In practical embodiments of theencoder spectrograph100, it may be useful to consider optical elements (e.g., mirrors, grating substrates, and lenses) that have surfaces (i.e., the surface sag, not to be confused with the shape of the aperture) that are rotationally symmetric as opposed to biconic or torroidal surfaces, and lenses that are spherical rather than aspherical. Although embodiments of theencoder spectrograph100 incorporating non-rotationally symmetric optical components and aspherical lenses may demonstrate higher optical performance, cost and manufacturability considerations may drive the decision for what is best suited for a given application. In one practical embodiment ofencoder spectrograph100, thediffraction grating102 is formed using a holographic etching process on a spherical substrate having a radius of curvature of roughly 100 mm, the crossover-focusingoptic103 is an asphere having a nominal radius of curvature of roughly 400 mm, the crossoverfast collection optic105 is an asphere with a nominal radius of curvature of roughly 30 mm, and thedetector lens106 has a spherical surface with a nominal radius of curvature of 30 mm when fabricated from silicon or other material with a similar index of refraction.

Mounting Structure for Encoder Spectrograph

FIG. 4 is an exploded view of a dual-spectrograph encoder200, in accordance with an embodiment of the invention. In the dual-spectrograph encoder200, two optical systems are contained within the same device and can be used to encode two independent dispersed images simultaneously using a single encoder disc. Each of the optical systems in theencoder200 may comprise the double-crossover encoder optics100 shown inFIGS. 1-3. Accordingly, like reference numerals are used to show the correspondence between the optical components shown inFIGS. 1-3 and the physical components shown inFIGS. 4-6. As illustrated inFIG. 4, an embodiment of a dual-spectrograph encoder200 implements a modular bracketing architecture to mount the optical components of theencoder spectrograph100. InFIGS. 5 and 6, the two spectrographs are designated as spectrograph A and spectrograph B. For brevity, one of the spectrographs is described in detail in the description of the modular bracket structure shown exploded inFIG. 4.

Thespectrograph200 receives light through anentrance aperture101. Theentrance aperture101 comprises an aperture in a thin foil that is mounted onto anentrance aperture bracket201. In this manner, the optical resolution of thespectrograph200 can be varied by interchanging foils with different aperture widths. In one embodiment of dual-spectrograph encoder200, the two aperture foils are different.

The grating102 is mounted into agrating mount202. Thegrating mount202 includes a mounting surface, receptacle, or cavity that incorporates a compound angle. The compound angle is defined by the inclination angle θ_incand an auto-collimation angle θ_auto, as shown inFIG. 1 andFIG. 2, respectively. In this manner, thesystem200 can be configured operate in different spectral ranges by interchanging a system of gratings designed for the same grating mount, but having different grating patterns optimize for different spectral ranges.

Anoptics bracket203 includes mounting provisions (including angled surfaces) to hold and position thecrossover focusing reflector103, thefast collection optic105, and thedetector lens106. Theoptics bracket203 also includes a detector insertion receptacle210 (e.g., a TO-8 compatible) to facilitate the mounting and alignment of the detector209 (which comprises thedetector window107 and the detector element108) to the other optical elements. The detector209 is held into thedetector insertion receptacle210 by adetector backing plate211, which ensures good thermal contact between the detector209 and theoptics bracket203. In this manner, theoptics bracket203 adds thermal mass to detector209 to improve the detector's thermal stability—useful, for example, for infrared applications.

Anencoder baseplate204 provides a common mounting surface to locate precisely theentrance aperture bracket201, thegrating mount202, theoptics bracket203, and aspindle motor242 relative to one another. This simplifies alignment and assembly of thesystem200. Aspindle motor242 is mounted onto the top side of a baseplate204 (i.e., the spectrograph side), and theencoder104 is mounted into a recessed cavity located on the bottom side ofencoder baseplate204 usingspindle hub243. Also shown inFIG. 4 is arotary encoder244, which provides signals to synchronize the data acquisition from the detector209 to the rotation angle of theencoder disc104.

Theoptics bracket203 is mounted onto theencoder baseplate204 so that the dispersed image I_disis projected through a field-of-view aperture111 onto theencoder104. The field-of-view aperture111 serves to control the illuminated width of the dispersed image along the azimuthal axis of theencoder104 to match the half-tone area used in the design of the radiation filter as described in U.S. Pat. No. 6,995,840. In addition, field-of-view aperture111 serves to limit the surface area of theencoder104 that is visible to the detector109 to minimize effects of modulated background blackbody radiation on the detector signal.

The location ofspindle motor242,grating mount202 and optics mount203 on the top surface ofencoder baseplate204 minimizes or reduces the volume of theencoder spectrograph200. Other arrangements that locate the spindle motor on the bottom side ofencoder baseplate204 may be employed.

Also shown inFIG. 4 is detector preamplifier printedcircuit board212, which is mounted onto a backdetector backing plate211. The preamplifier printedcircuit board212 includes a clear aperture to allow the incident radiation beam to pass through the plane of a printedcircuit board212 betweenentrance aperture101 and grating102 without obstruction. Other shapes for detector preamplifier printedcircuit board212 used to prevent the obstruction of the incident radiation beam may include different cutouts or apertures. The location and orientation of the printedcircuit board212 can be chosen to mitigate potential sources of noise in the detector signal path by minimizing the distance between the signal leads of the detector209 and the input to the gain stage of the preamplifier circuit (contained within the printed circuit board212). Alternatively, the printedcircuit board212 may be located outside of the optical path.

In addition, atemperature sensor213 may be mounted to theencoder baseplate204 to provide a measurement of the temperature of the optical system. Preferably, the data provided by thetemperature sensor213 is used in one or more compensation algorithms to correct or provide context to the signals provided by the detector209.

Also shown inFIG. 4 is the optics enclosure comprising,enclosure housing280,enclosure coverplate gasket281,enclosure coverplate282,enclosure mounting flange283, enclosure windows284 (A and B), enclosure window o-rings285 (A and B), and enclosure window retainer plates286 (A and B).Encoder baseplate204 drops into and mounts into the bottom ofenclosure housing280 to provide mechanical and dust protection and an increased thermal mass for the entire optical system. Preferably, the optical enclosure includes provision for gas purging286, and cable feed-thrus287 with grommets (not shown) to provide a significantly air-tight enclosure. In this manner, the effects of background chemical concentrations on the uncontrolled path (i.e., the optical path outside of any sampling system) of the optical system can be mitigated.

In one embodiment, thesystem200 is readily configured for different spectral ranges by interchanging thegrating102 and the detector109 with different gratings and different detectors optimized for different spectral regions. For example, any detector available in a standard package (e.g., a TO-8 package) may be inserted into thedetector insertion receptacle210, and a continuum of spectral regions can be imaged onto theencoder104 by varying the grating pattern on the substrate of thegrating102. Smaller detector packages can also be made compatible with thedetector insertion receptacle210 using an adaptor ring.

For a system spanning the 2.5 to 5-micron spectral range, the optimized auto-collimation angle (θ_incinFIG. 1) may be about 10 degrees. The optimized auto-collimation angle is a function of the free spectral range multiplier, which is defined as the ratio of the longest wavelength to the shortest wavelength of the spectral range imaged on theencoder104 in a given diffraction order. For example, in spectrographs spanning 2.5 to 5 microns, 5 to 10 microns, and 7 to 14 microns, the ratio of the longest wavelength to the shortest is 2×. For a 2× free spectral range multiplier embodiment, the optimized auto-collimation angle is about 10 degrees, whereas for the 3 to 5-micron system described above (5/3× free spectral range multiplier), the optimized auto-collimation angle is about 13 degrees. Preferably, the free spectral range multiplier is less than or equal to 2×. A spectral range of 2.5 to 5.5 microns is chosen to illustrate one embodiment of the invention, and other spectral ranges may be chosen, including 1.375 to 2.75 microns, 2.75 to 5.5 microns, 6 to 12 microns, and 7 to 14 microns.

In a group of instruments spanning different spectral ranges but having a common free spectral range multiplier, a singlegrating mount202 can be used to mount the respectivedifferent diffraction gratings102. Modular bracketing may be used to simplify the manufacturing of various spectral analysis instruments based on theencoder spectrograph200.

Dual Encoder Spectrograph

FIG. 5 illustrates an assembled view of the embodiment of the dual-spectrograph encoder200 ofFIG. 4. In this embodiment, asingle encoder disc104 and asingle spindle motor242 are used to encode dispersed radiation simultaneously using a pair of spectrographs, A and B. In one embodiment, each spectrograph A and B implements theoptics system100 described inFIGS. 1-3. As illustrated, for each spectrograph A and B there is an entrance slit101-A and101-B mounted on acommon bracket201. As shown inFIG. 5, each spectrograph A and B of the dual-spectrograph instrument200 includes a grating mount202-A and202-B that secures the corresponding diffraction grating,102-A and102-B, respectively, and an optics bracket203-A and203-B that holds the corresponding crossover focusing mirror103-A and103-B, the crossover fast-collection optic105-A and105-B, the detector lens106-A and106-B, and the detector209-A and209-B, respectively.

As shown inFIG. 5, thespindle motor242 is located on the same side and sandwiched between the encoder spectrographs A and B. In one embodiment, optics brackets203-A and203-B have machined recesses (shown inFIGS. 5 and 6) to enable a closer placement to the spindle motor. In this manner, the overall size of theinstrument200 is made very compact and has a relatively small footprint. Thespindle motor242 can also be located on the opposite side of the baseplate205.

FIG. 6 is a plan view of the double-crossover encoder spectrograph200 shown inFIG. 4. As illustrated, radiation beams from two independent external sources224-A and224-B are imaged through corresponding samples226-A and226-B and onto corresponding entrance apertures101-A and101-B. The corresponding diffraction gratings202-A and202-B produce corresponding dispersed images at two different locations on the surface of theencoder disc104 after passing through the respective encoder apertures111-A and111-B. As theencoder disc104 rotates, the dispersed images thereon are simultaneously encoded to provide two independently encoded radiation beams, which are focused onto the detectors209-A and209-B. The signals produced by the detectors209-A and209-B are then provided to and analyzed by a computer to determine the spectral properties of each of the radiation sources224-A and224-B. In this manner, asingle instrument200 having a single encoder disc104 (i.e., a single moving part) can be used to analyze two independent radiation sources224-A and224-B. In this way, asingle encoder disc104 and asingle spindle motor242 are used to analyze radiation transmitted through two different sources224-A and224-B or to analyze radiation transmitted through acommon sample226 and/or with different path lengths.

In an alternate embodiment, additional external optics are used to split radiation from a single source into two beams to provide input radiation for the encoders A and B. In this manner, asingle instrument200 having a single encoder disc104 (i.e., a single moving part) can be used to analyze simultaneously asingle radiation source224 in two independent spectral ranges or to analyze simultaneously asingle radiation source224 altered by two distinct samples226-A and226-B.

Preferably, the diffraction gratings202-A and202-B, and the detectors209-A and209-B are selected and/or optimized for the analysis of their corresponding radiation sources224-A and224-B (or samples226-A and226-B). For example, the radiation sources224-A and224-B (or samples226-A and226-B) may have corresponding spectral features in different wavelength ranges that require different grating periods and/or detector types for detection and analysis. In such a case, different diffraction gratings202-A and202-B would be used where each grating202-A and202-B is tailored for the source224-A and224-B that is to be encoded.

InFIG. 6, radiation provided by the source224-A and transmitted through the sample cell226-A (e.g., a first liquid cell, gas cell, or ATR) is imaged on the corresponding entrance aperture101-A, and radiation provided by the source224-B and transmitted through the sample cell226-B (e.g., a second liquid cell, gas cell, or ATR) is imaged on the corresponding entrance aperture101-B. In this manner, a single instrument having a single encoder disc can be used to analyze two independent samples226-A and226-B.

If the sample is a gas, vapor or liquid, it may be advantageous to provide samples contained that have substantially identical constituents in substantially identical concentrations to facilitate the analysis of radiation transmitted through a common sample using different optical path lengths. For example by connecting the samples226-A and226-B with pneumatic tubing a continuous flowing sample (e.g., natural gas) can be simultaneously analyzed (compensating for flow latency, if required) using a short path cell to quantify the stronger absorbing (e.g., higher concentration) constituents and a long path cell to quantify the weaker absorbing (e.g., lower concentration) constituents. In this manner, the chemical concentration dynamic range is enhanced relative to instruments using a single sample cell with a fixed length.

Preferably, the radiation sources224-A and224-B, the sample cells226-A and226-B, the diffraction gratings102-A and102-B, and the detectors209-A and209-B, are selected and/or optimized for the analysis of the corresponding samples226-A and226-B, respectively. InFIG. 6, for example, the sample226-A is depicted as providing a much shorter path length than the sample226-B. The different sample size may be useful in situations where samples have constituents in wide-ranging concentrations (e.g., natural gas may have 85% methane and <1% hexane) or where samples of mixed phases (e.g., solid, liquids and gasses) are being analyzed. The illustration of transmission cells inFIG. 6 was chosen for clarity. Other sampling interfaces may be employed that utilize transmitted, reflected, scattered, or emitted radiation from samples in solid, liquid, gas, and/or vapor phases.

Temperature Compensation by Selecting Sign of Auto-Collimation Angle

The radius ofencoder disc104 and the nominal grating period ofdiffraction grating102 increases with increasing temperature. As a consequence, the diffraction angle (as measured from the non-diffracted radiation beam) of a given wavelength component decreases with increasing temperature. It is advantageous to select the sign of the auto-collimation angle (i.e., diffraction order) such that zeroth order (non-diffracted) radiation is directed away from the center of the encoder disc. As temperature increases, the period of thediffraction grating102 is reduced due to the expansion of the substrate material. This, in turn, moves the diffracted spectral components along the negative Y-axis (ofFIG. 2) toward the zeroth-order or non-diffracted radiation beam. As temperature is increased, the radius of theencoder disc104 increases due to the expansion of the substrate material, which moves the encoding tracks away from the axis of rotation (i.e., the center of the encoder disc104).

To compensate for these temperature effects, in one embodiment, the sign of the auto-collimation angle is selected so that the two temperature effects compete against each another and, preferably, cancel each other out. Preferably, the sign of the auto-collimation angle of thediffraction grating102 is selected so that the zeroth-order (i.e., non-diffracted) radiation is directed away from the disc center (i.e., axis of rotation). In this configuration, the change in the center wavelength of a given encoded component as a function of temperature is minimized.

Applications

A dual-spectrograph system operating in the Near Infrared (NIR) (e.g., using InGaAs, InAs, PbS, PbSe, thermopile or pyroelectric detectors) and Mid Infrared (MIR) (e.g., using PbSe, InSb, HgCdTe, thermopile or pyroelectric detectors) enables the consumer to compare in real-time chemometric analyses in two distinct spectral ranges, or to augment one with the other for improved reliability, dynamic range and/or accuracy. The dual-spectrograph system may be implemented using the dual encoder spectrograph architecture described herein. Other embodiments using other combinations of spectral ranges (e.g., ultra-violet (UV)-NIR and UV-MIR) may be used.

In another embodiment, a single instrument uses two spectrographs operating in different and distinct spectral regions (e.g., NIR and MIR), preferably mounted onto a single encoder. A sampling interface with two sampling sub-systems optimized for liquid (and/or solid) and gas (and/or vapor) phase samples. Each sampling sub-system is interfaced with one of the spectrographs to enable the simultaneous analysis of samples having both liquid and solid, liquid and gas (or vapor), and/or solid and gas constituents (i.e., mixed phase systems). Preferably, the sampling sub-systems (e.g., path length, temperature, and the like) and the corresponding spectrographs (e.g., spectral region, resolution, selection of encoded components, spectrum acquisition rate, and the like) are optimized for the analysis of the mixed-phase systems. For example, in a Liquid+Vapor Analyzer, the path length for the liquid sampling sub-system may be shorter that the path length for the vapor sampling sub-system, and/or the NIR spectral region may be used for the liquid analysis, and the MIR spectral region may be used for the vapor analysis. Such an instrument can be used for analyzing both the liquid and the headspace (i.e., the area above the liquid containing gasses or vapors) in fermentation vats, drums, solvent dryers, and other mixed-phase systems. This application can be implemented with a dual-spectrograph system as described herein.

In another embodiment, a single instrument uses two spectrographs interfaced with two sampling sub-systems: (1) an Attenuated Total Reflectance (ATR) system for analyzing solids (e.g., powders) and/or liquids; and (2) a gas cell for analyzing ambient air. In this manner, a first responder system can analyze samples at the scene of an incident while simultaneously monitoring the ambient environment for potential hazards (e.g., toxic and explosion hazards), all within a single instrument. This application can be implemented with a dual-spectrograph system as described herein.

Spectral-Library Interface for EPIR Analyzer

One challenge in the deployment of a new type of spectral instrument is the ability to share spectral information and compare results between two or more instruments. Because of manufacturing tolerances, the spectra produced by two encoder spectrographs are generally not expected to be identical. As such, algorithms are used to translate spectra for comparison purposes. Instrument specific spectral parameters (ISSPs) are the parameters that are used by such algorithms to enable spectra to be translated back and forth between two or more instruments.

Standard high-resolution (e.g., FTIR) reference spectra may be translated (e.g., masked and deresolved) using a model for the encoder-spectrograph's instrument specific spectral parameters (ISSPs) (e.g., dispersion function and radiation filter configuration) to provide instrument-specific reference spectra (ISRS) that are fit to “as-measured” spectra to determine the ISSPs for the encoder spectrograph. The spectral parameters enable the precise specification of the center wavelength and bandwidth for each encoded spectral component.

Reference spectra obtained from a spectral library may be processed (e.g., convolved) with the instrument's unique spectral parameters to provide instrument-specific reference spectra (ISRS). These ISRS are used with (e.g., compared to) “as-measured” spectra for chemometric analyses, e.g., a spectral library search to determine the origin of an unknown spectrum.

Existing calibrations can be transferred to the present invention by transforming high-resolution (e.g., FTIR) calibration spectra into instrument-specific calibration spectra (ISCS). The ISCS are then processed in the same manner as the original calibration spectra to generate an instrument-specific calibration (ISC). This process can be repeated on a number of instruments using their respective ISSPs to provide unique ISCs. In this manner, an existing calibration can be transferred onto one or more instruments.

In one embodiment, an algorithm that is compatible with the MyInstrument interface protocol, or other common data transfer or exchange format, inputs ISSPs and reference spectra from a spectral library and outputs provide instrument-specific reference spectra (ISRS) for use in the spectral analysis of an unknown sample.

In-Situ Phase Analysis for EPIR Analyzer Employing Orthogonal Encoded Components Having Substantially Identical Modulation Frequencies

U.S. Pat. No. 6,995,840 describes a modulator that provides pairs of encoded components having the same modulation frequency that are substantially orthogonal to one another. Since the maximum number of encoded unique harmonic components (i.e., each component encoded with a different harmonic of the fundamental rotation frequency of the modulator104) is limited by the dispersed image width and the circumference of the outermost encoding track on the modulator, this approach enables the multiplexing of up to twice as many encoded components. One approach described in U.S. Pat. No. 6,995,840 uses pairs of sinusoidal modulations that have the same frequency, but are 90 degrees out of phase (i.e., sines and cosines). The orthogonality of sine and cosine components having the same modulation frequency is limited by the uncertainty of the phase. It is therefore desirable to minimize the phase error to maximize inter-channel orthogonality.

In one embodiment, to minimize phase error and maximize orthogonality, an in-situ phase analysis based on prime-number encoded components is used. In a first step, two or more phase calibration components (PCCs) are provided with sine-only or cosine-only modulations. These PCCs are preferably prime-number harmonics of the fundamental rotational frequency, which are expected to have the least interference or cross-talk. The PCCs are then analyzed (e.g., using quadrature) to determine the phase of the encoded signals. A fit is generated between the resulting PCC phase and the modulation frequency and/or modulator radius using an appropriate model (e.g., a polynomial in modulation frequency and/or modulator radius). The resulting phase verses modulation frequency and/or modulator radius model is then applied to the other encoded components to maximize the orthogonality of sine and cosine components having the same modulation frequency.

TABLE 1
Harmonic	Phase	R0	ΔR	PCC

23	0	33.6932	0.0404	yes
24	0	33.774	0.0404	no
24	90	33.8549	0.0404	no
25	90	33.9358	0.0404	no
25	90	34.0166	0.0404	no
26	0	34.0974	0.0404	no
26	90	34.1782	0.0404	no
27	0	34.259	0.0403	no
27	90	34.3398	0.0403	no
28	0	34.4205	0.0403	no
28	90	34.5013	0.0403	no
29	0	34.582	0.0403	yes

Table 1 describes a portion of an embodiment of theencoder disc104 that uses 144 different harmonics to encode 256 different spectral components, to illustrate an embodiment of the invention where 32 prime number PCCs are used to provide input for the phase calibration algorithm. The table illustrates a case in which seven different harmonics are used to encode twelve different spectral components at twelve different radii on the surface ofencoder disc104. The table entries correspond to the (sinusoidal) radiation filter harmonic, the patterned phase, radial position (R₀), radial width (ΔR), and whether the encoded component is a PCC. As shown in the table, the encoded components that are not designated as PCCs are used in pairs having the same modulation frequency (i.e., harmonic of the rotation frequency of encoder disc104), which are modulated at 0 and 90 degrees, respectively. In contrast, the PCCs are not paired and are all modulated at a phase of 0 degrees. To assign the correct proportion of the amplitude to the two encoded components comprising the pair of components having the same modulation frequency (i.e., the non-PCCs), the phase of the signal as measured by the detector209 at the modulation frequency of the pair must be known.

The phase as measured by the detector209 (signal phase) is different from the patterned phase of theencoder disc104 due to the frequency response of the detector electronics and because of non-lineal aberrations of the dispersed image along the radial axis ofencoder104. In one embodiment, to mitigate these uncertainties in the signal phase and recover the correct proportion of the amplitude to the two encoded components comprising the pair of components having the same modulation frequency, the phases of the PCCs are determined and an interpolation is used between the PCCs to estimate the signal phase of the non-PCC components.

In Table 1, for example, the signal phase of the 23rd harmonic and the 29th are determined by taking the arctangent of the sine and cosine projections of these harmonics relative to a global reference phase (e.g., as defined by an analysis of one or more of the signals provided by therotary encoder244 or a persistent feature in the detector signal, such as a centerburst). Once the signal phase of the 23rd and 29th harmonics have been determined, the signal phase of the components of the 24th through the 28th harmonics that have the same patterned phase as the PCCs (i.e., 0 degrees in Table 1) can be estimated by interpolation (e.g., using the radial position and/or the modulation harmonic as the abscissa). The corresponding signal phases for the 24th through the 28th harmonic components that have patterned phase different from the PCCs (i.e., 90 degrees in Table 1) are determined by interpolation and adding or subtracting the phase difference between the component and the PCC (e.g., 90 degrees). Once the signal phases of the 24th through the 28th harmonic components have been estimated, the proper proportion of the amplitudes at a given harmonic can be assigned to the two out-of-phase components using the corresponding trigonometric projections.

In another embodiment, the phase calibration described above is performed as a background task, at periodic intervals, or in response to external triggers (e.g., a change in the temperature as measured by temperature sensor213). Preferably, the phase calibration algorithm uses the PCC signal amplitudes to weight the importance of the PCCs in the fitting routine. In this manner, a poor fit resulting from a strong absorbance in one or more of the PCCs can be prevented.

Spectrometer Calibration Training Mode Employing Unique Spectra Sorting Algorithm

One challenge in developing a new chemometric application to identify and quantify analytes by analyzing unknown spectra is acquiring the proper reference spectra required for the underlying calibration. The calibration training spectra are the set of spectra required to build a calibration (e.g., construct a complete set of variance spectra) for a given chemometric application. There are many recommendations in the literature for selecting the calibration training spectra, but from a mathematical perspective, the requirement is simply to include all anticipated unique spectra—i.e., spectra that are not simply related to one another by an amplitude scale factor in the absorbance regime (or equivalently, a power factor in the transmission regime) or are simply linear combinations of two or more calibration training spectra.

An embodiment of the invention enables user to collect the calibration training spectra without having to follow a complicated script.FIG. 7 illustrates a method of determining calibration training spectra for a spectrometer, in accordance with an embodiment of the invention. The embodiment analyzes701 a continuous stream of spectra acquired during a calibration training mode and automatically logs702 those unique spectra required to build chemometric calibration model (hereafter referred to as “calibration training spectra”). A unique spectra sorting algorithm (USSA) determines if a new spectrum is unique within the growing set of calibration training spectra by comparing703 the new spectrum to each of the existing calibration training spectra (e.g., by fitting the new spectrum to each of the calibration training spectra using an amplitude scaling parameter in the absorbance regime or a power scaling parameter in the transmission regime). The algorithm then calculates704 a residual spectrum by subtracting the scaled new spectrum from the existing calibration training spectrum. A residual score is calculated705 for the new spectrum for each of the existing calibration training spectra (e.g., by comparing the integrated absorbance of the residual spectrum to a user-defined threshold). The residual scores are sorted to identify706 the minimum residual score, which identifies707 the calibration training spectrum that is the closest match to the new spectrum.

If the minimum residual score is determined to be significant, the new spectrum is deemed to be unique and is added to the set of calibration training spectra. If the spectra is determined not to be unique but corresponds to a higher amplitude (or concentration) version of an existing spectrum (e.g., the absorbance amplitude scaling parameter is greater than 1), the new spectrum overwrites the existing unique spectrum. In this manner, spectra with the highest signal-to-noise ratio are retained in the set of unique calibration training spectra.

The term spectrum as used herein has a broad definition, and the term includes selected portions of the entire spectrum measured by the instrument. In this manner, spectra from non-overlapping analytes can be separated from one another in mixtures to prevent unnecessary growth in the number of calibration training spectra.

In one embodiment, a graphical user interface (GUI) enables the user to “tag” spectra with analyte species and concentration information. Preferably, the GUI allows the user to tag spectra during the collection of the calibration training spectra and presents the user with a summary graph of the amplitude of the unique spectra versus time at the conclusion of the training session. This allows the user to verify and correct existing tags or add additional tags to the training set. In another embodiment, responsive to a user click on the graph the GUI presents a graph of one or more selected unique calibration training spectra, which can then be compared with spectra from a spectral library or other database for further verification of assigned tags.

In another embodiment, the calibration training spectra are stored in a database within the instrument that produced them, and the internal database are synchronized with or otherwise transferred to an external database (e.g., via the Internet) to enable access to the calibration training spectra for the development of chemometric calibrations. Preferably, the externally developed chemometric calibration is then transferred into the instrument's internal database for an appropriate chemometric application.

Summary

While the invention has been described above by reference to various embodiments, it will be understood that different combinations, changes and modifications may be made without departing from the scope of the invention, which is to be defined only by the appended claims and their equivalents.

For example, instead of using the specific optical elements in the specific order as described, including the placement of a sample cell, or sample collection in the beam path, other optical elements, optical systems, or arrangements may be used without departing from the scope of the invention. Moreover, a transmissive spatial light modulator can also be used with the encoder spectrograph with appropriate changes to the configuration of the optical elements. One embodiment of the encoder spectrograph can encode radiation in the 3 to 5-micron spectral range; however, persons of ordinary skill in the art will understand that with minor modifications (e.g., grating line spacing, holographic etching pattern, auto-collimation angle, detector lens material/coating, and/or detector type) the encoder spectrograph can be embodied to encode radiation in other spectral regions.

These and other variations are within the scope of the invention. Accordingly, the foregoing description of the embodiments of the invention has been presented for the purpose of illustration; it is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Persons skilled in the relevant art can appreciate that many modifications and variations are possible in light of the above teachings. It is therefore intended that the scope of the invention be limited not by this detailed description, but rather by the claims appended hereto.

Claims (9)

1. A method of determining calibration training spectra for a spectrometer, the method comprising:

analyzing a continuous stream of spectra acquired during a calibration training mode;

logging unique spectra to build calibration training spectra;

determining if a new spectrum is unique within the calibration training spectra by comparing the new spectrum to each spectrum of the existing calibration training spectra;

calculating a residual spectrum by subtracting the new spectrum from an existing calibration training spectrum;

calculating a residual score for the new spectrum for each spectrum of the existing calibration training spectra;

identifying a minimum calculated residual score from the calculated residual scores; and

identifying the spectrum of the calibration training spectra that is the closest match to the new spectrum by identifying the minimum calculated residual score.

2. The method ofclaim 1, wherein determining if a new spectrum is unique comprises fitting the new spectrum to each of the calibration training spectra using an amplitude scaling parameter.

3. The method ofclaim 1, wherein determining if a new spectrum is unique comprises fitting the new spectrum to each of the calibration training spectra using a power scaling parameter.

4. The method ofclaim 1, wherein calculating a residual score for the new spectrum for each of the existing calibration training spectra comprises comparing an integrated absorbance of the residual spectrum to a user-defined threshold.

5. The method ofclaim 1, wherein identifying a minimum score comprises sorting the calculated residual scores.

6. The method ofclaim 1, further comprising:

responsive to determining the new spectrum is unique, adding the new spectrum to the calibration training spectra.

7. The method ofclaim 1, further comprising:

responsive to determining the new spectrum is not unique, determining if the new spectrum corresponds to a higher amplitude version of an existing calibration training spectrum;

responsive to determining the new spectrum corresponds to a higher amplitude version of an existing calibration training spectrum, overwriting the corresponding existing calibration training spectrum with the new spectrum.

8. The method ofclaim 7, further comprising:

retaining the new spectrum in the calibration training spectrum.

9. The method ofclaim 1, wherein the continuous stream of spectra comprises selected portions of an entire spectrum measured by the spectrometer.

Images